SUMMARY REPORT on EPSRC grant GR/K04699 (1994-1997)
Engineering and Physical Sciences Research Council

Property-process-fibre distribution relationships
in fibre-reinforced composites

Principal Investigator: Dr John Summerscales
Department of Mechanical and Marine Engineering
University of Plymouth

Co-Investigator: Dr D E Wright (UoP Department of Biological Sciences)


Digitally enhanced partial cross-section
of the composite micrograph for
a flow-enhanced reinforcement fabric

The EPSRC research grant was awarded to Dr Summerscales and Dr Wright for a 3-year programme employing a researcher studying for a PhD degree, with technician assistance. A complementary grant (EPSRC GR/J77405) was awarded to Dr Felicity Guild (Universities of Surrey/Bristol) to permit collaboration and attendance at conferences.

Resin transfer moulding is a cost-effective processing method for composites limited by the time taken to fill the mould. Specialist commercial fabrics expedite the flow of the resin by clustering the reinforcement fibres. However, such clustering leads to significant areas of 'pore space' (resin-rich areas in the laminate). These resin-rich areas have been shown to lead to a deterioration in mechanical properties, both from predictive analyses and by experimental measurements. It was the overall objective of this work to address this process-property dilemma.

The main objectives of the proposed research were:

  • development of a quantitative description for the fibre distribution of fabric reinforced composites
  • measurement of laminate properties as a function of fibre distribution
  • correlation of theory and experimental measurements on laminates with different fibre distributions
  • development of property-process-fibre distribution relationships

Three CFRP materials were studied: composites based on twill fabrics from Griffin's doctoral research are referred to as 'Griffin/Carr fabrics', 'Brochier fabric' laminate materials from EU grant BE5477 and novel flow enhancement fabrics woven by Carr Reinforcements (SME based in Stockport) termed new Carr fabrics.

Initial training of the RA was undertaken using the Griffin/Carr materials. Values of total area of pore space arising from areas in the range 0.1-0.25 mm2 were correlated to the measured permeability.

The main work was conducted on Brochier fabrics (2x2 twill, 5H-satin and Injectex 5H-satin) of equal fabric areal weight produced from the same batch of fibre. The Injectex flow-enhanced satin had a significant number of large pore space areas (in the range 0.08-0.3 mm2). The use of cumulative feature frequencies to describe the microstructure is very dependent on subjective interpretation of the histograms presented. Fractal dimension: the slope of the plot of detected area versus detection box area on a log-log (Richardson) plot was chosen to characterise the new Carr fabrics. The parameter has not previously been used for continuous fibre composites.

Mechanical testing was carried out: interlaminar shear strength to CRAG 100, tensile modulus and strengths to CRAG 302 and compressive modulus and strengths to CRAG specification 401. When the experimentally measured properties for Brochier fabrics are ranked from lowest to highest value, each triad of data points (one per fabric) is ranked in the same sequence for all ten samples for each of the five mechanical properties tested. The fractal dimension (slope of the Richardson plot) is ranked in the same sequence as the mechanical properties (except ILSS) for four key parameters of the Brochier fabrics.

Table 1: Summary of results for all tests showing ranking for Brochier fabrics
PropertyRanking
Permeabilitysatin<Injectex<twill
ILSSsatin>Injectex>twill
Tensile strengthsatin>twill>Injectex
Compressive strengthsatin>twill>Injectex
Tensile modulussatin>twill>Injectex
Compressive modulussatin>twill>Injectex
Fractal dimensionsatin>twill>Injectex

The new Carr fabrics contain 33% (A), 20% (B), 14% (C) and no (D: reference material) flow-enhancing tows.

Table 2: Summary of results for all tests showing ranking for new Carr fabrics
Note that the sequence is not correlated directly to the proportion of flow-enhancing fibres.
PropertyRanking
PermeabilityA>C>B>D
Fractal dimensionA>C>B>D

The completion of the mechanical testing for the new Carr fabric laminates has been delayed. This should be complete by mid-October, with analysis completed this calendar year and a conference presentation in March 1999.


REPORT FOR 'DIPSTICK' MEETING: Engineering and Physical Sciences Research Council grant GR/K04699

Property-process-fibre distribution relationships in fibre-reinforced composites

Principal Investigator: Dr John Summerscales
Department of Mechanical and Marine Engineering
University of Plymouth (UoP)
Co-Investigator: Dr D E Wright (UoP Department of Biological Sciences)

Resin transfer moulding is a cost-effective processing method for composites limited by the time taken to fill the mould. Specialist commercial fabrics expedite the flow of the resin by clustering the reinforcement fibres. However, such clustering leads to significant areas of 'pore space' (resin-rich areas in the laminate). These resin-rich areas have been shown to lead to a deterioration in mechanical properties, both from predictive analyses and by experimental measurements. It was the overall objective of this work to address this process-property dilemma.

The Final Report on this EPSRC research grant was submitted at the end of September 1998. The completion of the mechanical testing for the new Carr fabric laminates (based on a 2x2 twill weave) was delayed and completion was anticipated by mid-October, with analysis completed this calendar year and a conference presentation in March 1999.

Mechanical testing was carried out: tensile modulus and strengths to CRAG 302 and compressive modulus and strengths to CRAG specification 401. The mechanical testing of the strain-gauged specimens is now complete. Summary results are attached for the secant moduli at 2500 (( (Chart 2) and failure stresses (Chart 3) in both tension and compression for each of the warp and weft directions of the Carr fabrics.

The fabrics are designated as follows according to the number of flow enhancing tows (FET):

Carr noRefFETFabric description
16Dzerostandard fabric used as a reference material
15C14%one flow enhancement tow to every six standard tows in the weft direction
14B20%one flow enhancement tow to every four standard tows in the weft direction
13A33%one flow enhancement tow to every two standard tows in the weft direction
13aX66%two flow enhancement tows to every one standard tow in the weft direction

The use of cumulative feature frequencies to describe the microstructure is very dependent on subjective interpretation of the histograms presented. Fractal dimension: the slope of the plot of detected area versus detection box area on a log-log (Richardson) plot was chosen to characterise the new Carr fabrics. The parameter has not been used for continuous fibre composites prior to this project.

Table 1: Summary of results for all tests showing ranking for new Carr fabrics
Note that the sequence is not correlated directly to the proportion of flow-enhancing fibres.
PropertyRanking
PermeabilityA > C > B > D
Fractal dimensionA > C > B > D
Compression modulus (warp)D ~ C ~ B ~ A
Tensile modulus (warp)D ~ C ~ B ~ A
Compression modulus (weft)D > C > B > A
Tensile modulus (weft)D > C > B > A
Compression strength (warp)D ~ C ~ B ~ A
Tensile strength (warp)D ~ C ~ B ~ A
Compression strength (weft)D ~ C ~ B ~ A
Tensile strength (weft)D > B > C > A

The permeability ranks with fractal dimension, whilst the tensile strength in the weft direction (parallel to the flow enhancing tows (FET)) ranks with the inverse of the fractal dimension. Otherwise, the mechanical properties are either independent of the FET, or decrease as the proportion of flow enhancing tows is increased. Tensile moduli are always greater than compressive moduli, as would be expected when straightening the fibre crimp in tension rather than exaggerating the pre-buckling in compression. The compressive failure strengths are independent of the test direction, which would be consistent with the higher crimp associated with crossing normal tows than when crossing the smaller FET.